Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
NCTM or The Math Forum.



Simple(?) chisquare question
Posted:
Aug 6, 1996 10:24 AM


Hi,
I'm having a dispute with a friend over chi square statistics, and he refuses to believe he could be wrong.
He has a model for predicting reaction times and compares these directly with the measured reaction times and does a chi square on them. E.g.
MODEL DATA 4 2 6 4 8 10
and he actually computes chi2 = (42)^2/2 + (64)^2/4 + (810)^2/10) = 3.4 and then looks it up in the table and yells "SIGNIFICANT".
I try to tell him that this method of operation is illegal. E.g., if you express your reaction times in seconds or milliseconds you get a great difference in the outcome as you can imagine.
Does anyone have an argument or a reference that will convince either him or me that he is/I am wrong?
Thanks!
 Theo Vosse  Unit for Experimental Psychology University of Leiden The Netherlands



